Today, companies have more information at their fingertips than ever before.
In the world of 2024, good data is critical to making the right decisions. It quite often sits at the base of a productive and responsive supply chain and is a key component to whether an organisation succeeds or not. Data doesn’t just provide insights, it encourages teams to eliminate manual work to deliver efficiency and accuracy which is also being amplified through the likes of generative AI. Truthfully, the supply chain space has never seen so much transformation before and the real winners will be the ones who steer into this technological evolution and are open to embracing new solutions quickly.
Following the recent ASCM CONNECT, Matt McKinney, Co-founder and CEO of Loop spoke to CPOstrategy, to discuss the importance of using AI to access better data, automate processes, and uncover insights in order to enable supply chain teams to maximise their people and performance.
Would you be able to give me a brief introduction to your role and the company you work for?
Matt McKinney: “I’m Matt McKinney, Co-founder and CEO of Loop. Loop is on a mission to unlock profit trapped in the supply chain and lower costs for consumers. Bad data and inefficient workflows create friction that limits working capital and raises costs for every supply chain stakeholder. Loop’s logistics-AI centralises and standardises supply chain and financial data to automate workflows and surface strategic insights. We improve visibility so companies can control costs, uplevel decisions, and power profit.”
ASCM CONNECT 2024
What is the value of events like ASCM CONNECT 2024: North America? How important is this conference in the supply chain calendar?
Matt McKinney: “ASCM is an incredible watering hole for an industry that is spread out. There is no one region that houses all of the supply chain companies so having the opportunity to come together and share insights and learn is really powerful.
“Because the supply chain is a connected network of networks – your success relies on the performance of your partners. Collaboration, trust, and transparency are paramount. Getting together at ASCM fosters all three and gives people the chance to understand innovation in the space.”
Global supply chain
Given the backdrop of the global disruption over the past few years (COVID, wars, inflation etc), how would you sum up where the supply chain space finds itself today?
Matt McKinney: “The logistics industry is in the midst of a major shake-up, driven by the rapid pace of technological change. As nearshoring, shifting customer expectations, regulatory changes, and a soft freight market weigh on businesses, companies are under pressure to keep up. Many leaders are turning to AI to revolutionise their supply chains. The challenge right now is that those that fail to adapt to these modern times risk falling behind.”
What do you feel are the biggest lessons supply chains have learnt over the past few years and how well equipped is the modern day supply chain now to deal with ‘black swan’ events like the ones we’ve seen recently?
Matt McKinney: “The disruptions over the last few years, especially COVID, exposed the companies that hadn’t invested in their people, processes, and technology. We’ve spoken to several teams who were stuck at home, unable to execute work without physically being in the office, because they hadn’t invested in software. Simultaneously, several teams were left in a lurch without great data to make informed decisions on how to best manage disruptions.
“As supply chains continue to get faster and more complex, teams need great data and automated processes so they can keep up and start maximising their personnel. Your people make your company. When people are struck processing, they cannot strategise and collaborate. Processing work hurts performance and often morale. As leaders, it’s our responsibility to give our people the tools to perform at their highest potential.
“AI is powering this transition. Sure, it’s not a magic wand, but it allows for a better understanding of your business and gets your people out of soul-sucking work that’s not strategic. This frees you to start thinking outside of your four walls and be more dynamic to optimise for any potential disruption.”
Digital transformation
Where are generative and conversational AI having the biggest impact in the supply chain? What types of supply chain and procurement decisions are now possible, or much easier to make, with the rise of AI and LLMs? How does this compare to just a couple of years ago?
Matt McKinney: “Generative AI foundation models, ChatGPT, Gemini, Claude, etc, have propelled AI into the mainstream. The reality is that AI has been around for a while but there has been incredible progress in two model types – Large Language Models and Computer Vision models that are at the forefront of transforming the AI ecosystem.
“The supply chain has struggled with the digital transformation because of its complexity and scale. The typical heuristics – ‘if this then that’ rules– of SaaS software can’t accurately handle the complexity, scale, and fragmentation of supply chain data. It’s a bad problem for people and legacy software but an incredible problem for AI.
“But the most important step is matching the right AI to the right problem. There are two core types of AI applications available today and which one you leverage will be dependent on the problem you’re trying to solve: General foundation models and domain-specific models.
“Now, general models like ChatGPT – GPT actually stands for General Processing Technology – are incredibly versatile. They’re our well-rounded models. They have a wide breadth of knowledge from history to pop culture to basic science. But that breadth is a double-edged sword when it comes to specialised business contexts.
“On the flip side, we have domain-specific models. These are our specialists, the experts in their field. In the supply chain world, we’re talking about models that are trained specifically on supply chain data, documents, and industry-specific terms.
“Now, why does this matter for supply chain? Well, imagine you’re trying to optimise your shipping routes or calculate complex transportation charges. General AI might give you a plausible-sounding answer, but it could be way off base. It simply doesn’t have the depth of knowledge required. That’s where our verticalised models come in.
“Verticalised models are built from the ground up to understand the nuances of supply chain language, operations, and context. They’re not just pulling from a general pool of internet knowledge. They’re working with curated, industry-specific data and ideally are trained on your data. This is logistics-AI.
“Supply chains generate vast amounts of data from diverse sources such as suppliers, customers, carriers, logistics service providers (LSPs), and internal systems. There is no ‘source of truth’ in the supply chain because each of these stakeholders has multiple systems with different IDs that track a component of their value delivery.
“Ask for a reference number on a shipment and you might get a number of different responses. Will you get the shipment ID? The purchase order (PO) number? Maybe it’ll be the carrier pro number? Or an invoice or order number?
“True logistics-AI is built to handle this unstructured data to create a unified view of your data from any source. At Loop, we use our models to wrangle and standardise supply chain data’s different taxonomies, terminologies, and discrepancies. So that we can improve visibility so companies can control costs, automate workflows, and uncover insights to power profit.
“The highest impact for AI in the supply chain will be on specialised AI that’s built for the supply chain. LLMs are language-based. The supply chain has its own language. To take advantage of AI, you need AI trained on logistics data, documents, and contracts. The beauty of a well-trained model is that it never stops learning and improving because it is constantly contextualising the inputs and outputs it can access. The more you use it, the more accurate and tailored to your business it becomes over time.
“Don’t get me wrong – general AI models have their place. They’re great for broad applications and simple tasks. But when it comes to the nitty-gritty use cases required in many use cases in the supply chain, domain-specific models are absolutely crucial. They’re the ones that can really drive efficiency, accuracy, and innovation in our industry.
“For a more deep dive to understand AI in the supply chain check out our free guide, Making Sense of AI In A Rapidly Evolving Supply Chain.”
GenAI journey
What should CSCOs out there do first? What needs to be considered before starting the gen AI journey?
Matt McKinney: “Choosing the right AI provider is about more than just selecting a vendor. It’s about applying the right AI to the right challenges AND finding a true partner who will grow with you and support your strategic goals. These are the key things that I would look for before bringing on AI:
“You need to understand if you need a generic or a specialised model. The more nuanced your problem or the more mission-critical it is (i.e. managing financial data), the more you’ll need AI that understands the context of your domain. We actually put together some questions you can ask to understand if a technology vendor offers generic or specialised AI, you can find them on our website: 10 Questions to Validate Logistics-AI.
“You should pick a partner, not a vendor. Select a provider who understands your business challenges and shares your vision for the future. This alignment is crucial for fostering innovation and ensuring that the provider can evolve alongside your business.
“You have to prioritize account support; you don’t want someone who is just going to leave it up to you after the sale or after the implementation. Partner with a team that is going to treat you as a top priority and deliver white glove support.
“Finally, you must assess viability. There are a lot of companies pitching AI solutions. Assess their long-term viability by looking at their growth, existing customers, and financial viability. Again, check out our great guide that dives deep into how to assess if an AI partner is the right fit for your supply chain needs.”
AI challenges
What are the biggest challenges or hesitations you’re seeing companies have around AI? What should organisations look for in technology to hedge against these concerns?
Matt McKinney: “Companies need to be very prescriptive about why they’re choosing AI. It’s a mistake to get the AI solution because it sounds good or seems like the right choice. Companies have sold ‘automation’ to everyone, but it comes in various forms. Automation can be a simple and rigid ‘if this, then that’ statements, it can be manual work executed by outsourced labor, or it can be true AI that is dynamic, flexible, and autonomously executing workflows…while continually learning, of course.
“The AI you need will be different based on the problem you’re trying to solve. Many supply chain problems require supply chain context. Generic AI models will not have the depth of knowledge or context needed to make high-quality decisions around the nuances of many supply chain use cases.
“Think about it, how contextual is your supply chain compared to your competitors or someone in a different industry? Your AI solution should be able to handle that because it’s trained only on your industry’s data, not trained on all of the data on the internet. That introduces a lot of risks you cannot control.
“I would recommend that the team be very clear about what their needs are before they start their research process. Then you need to ask very specific questions about how the provider’s AI works to understand what drives answers – an offshore team or actual models. Grill the providers a little bit. How do they train their AI, what data is it trained on, what is the rate of improvement, and what failsafes do they have to prevent hallucinations? All of these things are critical to ensure that the AI actually will provide the value the provider is claiming.”
What are the underlying issues in how companies are currently storing and looking at their supply chain data? Why is this a problem and how can they overcome those challenges with generative AI and knowledge graphs?
Matt McKinney: “Supply chains are drowning in data and it’s a mess—unstructured, fragmented, and disconnected. Legacy systems and outdated processes only make things worse. You’ve got rigid, siloed software—or no software at all, and people are handling data manually. You’ve also got a mixture of legacy documents (think paper invoices and photos of bills of ladings) and old-school filing systems that hurt decision quality. Combine that with the supply chain’s complexity, and it’s hard to get a complete view of the drivers that impact your profit. This locks up liquidity, slows down operations, and creates inefficiencies that no one can afford.
“But, AI thrives on messy, unstructured data. It turns this chaos into a powerful asset. In a world where manual processes and outdated tech are dragging you down, AI can boost efficiency, cut costs, and ramp up productivity. Logistics-AI makes data usable, it automates workflows and provides powerful analytics to improve margin. But again, AI is not a magic wand. It can certainly feel like one if you map the right AI to the right problem.”
Talent management
People are a company’s greatest asset but can also be a hurdle to overcome when it comes to innovation. How do you manage the people challenge and get them on board with change?
Matt McKinney: “Absolutely, people are a company’s greatest asset. You need to set your team up for success with the right technology, processes, and work. Historically, mind-numbing processing work has trapped people.
“The great thing is that AI takes this soul-sucking work off of people’s plates. One of our customers, Jeff Toman, a finance exec at Great Dane said it best, ‘Loop has turned my team from processors to analysts.’ Everyone wants to be an analyst. No one wants to be a processor. Jeff is empowering his teams to ask meaningful questions and make better decisions with high-quality data. This level of work elevates morale and builds confidence. Employees can see the impact of their work because it’s no longer just an operational slog. You get buy-in by showing them how their work will evolve and what else they’ll get to focus on with AI as their partner.”
Future
What about the future? How exciting a future does the supply chain space have?
Matt McKinney: “The supply chain is constantly changing. That’s why people like it because it keeps them on their toes. This is the most exciting time to be in the supply chain for two reasons. AI is helping customers accelerate their transformations so they can evolve from one phase of sophistication to another. We consistently help companies jump from Supply Chain 1.0 to 3.0, they completely bypass the challenges and investments required to get to 2.0. The rate of innovation has never been faster and it’s building a more transparent, collaborative, and performant supply chain.”
Find out more about Loop here.
- Digital Supply Chain
